Prior to 6.0, we relied completely on the publisher for EM, so the redundancy story was simple - everything on one node. Now that the Publisher is no longer a single point of failure for EM and login/logouts can be done on any node in the cluster, there are two things that must be considered for EM redundancy beginning with CM 6.0: IP Phone Services menu and the EM login/logout phone service.

IOS Server Load Balancing (SLB) - Where an IOS box front-ends CM servers such that all requests for phone services menu and EM login/logout service are sent to the IOS box which distributes the requests to the CM nodes in the cluster. This provides the ability to either load balance or bypass a node failure. Thus the Services URL for the IP Phone services menu (when you push the Services button on a phone) as well as the EM URL for the EM phone services are pointing to the IP/hostname of the IOS box. In this way, a failure of a node can be routed around on the fly and because in 6.0 EM logins are done at the local node, redundancies for all aspects of EM are possible.